Section 4.

4.3 Plane Axisymmetric Problems

In this section are considered plane axisymmetric problems. These are problems in which both the geometry and loading are axisymmetric.

4.3.1

Plane Axisymmetric Problems

Some three dimensional (not necessarily plane) examples of axisymmetric problems would be the thick-walled (hollow) cylinder under internal pressure, a disk rotating about its axis1, and the two examples shown in Fig. 4.3.1; the first is a complex component loaded in a complex way, but exhibits axisymmetry in both geometry and loading; the second is a sphere loaded by concentrated forces along a diameter.

Figure 4.3.1: axisymmetric problems A two-dimensional (plane) example would be one plane of the thick-walled cylinder under internal pressure, illustrated in Fig. 4.3.22.

Figure 4.3.2: a cross section of an internally pressurised cylinder It should be noted that many problems involve axisymmetric geometries but nonaxisymmetric loadings, and vice versa. These problems are not axisymmetric. An example is shown in Fig. 4.3.3 (the problem involves a plane axisymmetric geometry).

1 2

the rotation induces a stress in the disk the rest of the cylinder is coming out of, and into, the page

Solid Mechanics Part II

66

Kelly

Section 4.3

axisymmetric plane representative of feature

Figure 4.3.3: An axially symmetric geometry but with a non-axisymmetric loading The important characteristic of these axisymmetric problems is that all quantities, be they stress, displacement, strain, or anything else associated with the problem, must be independent of the circumferential variable . As a consequence, any term in the differential equations of 4.2 involving the derivatives / , 2 / 2 , etc. can be immediately set to zero.

4.3.2

Governing Equations for Plane Axisymmetric Problems

The two-dimensional strain-displacement relations are given by Eqns. 4.2.4 and these simplify in the axisymmetric case to

rr = r

u r r u = r r u 1 u = 2 r r

(4.3.1)

Here, it will be assumed that the displacement u = 0 . Cases where u 0 but where the stresses and strains are still independent of are termed quasi-axisymmetric problems; these will be examined in a later section. Then 4.3.1 reduces to

rr =

u r u , = r , r = 0 r r

(4.3.2)

It follows from Hookes law that r = 0 . The non-zero stresses are illustrated in Fig. 4.3.4.

Solid Mechanics Part II

67

Kelly

Section 4.3

rr

rr

Figure 4.3.4: stress components in plane axisymmetric problems

4.3.3

Plane Stress and Plane Strain

Two cases arise with plane axisymmetric problems: in the plane stress problem, the feature is very thin and unloaded on its larger free-surfaces, for example a thin disk under external pressure, as shown in Fig. 4.3.5. Only two stress components remain, and Hookes law 4.2.5a reads

rr =
with zz =

E 1 [ rr + ] rr = [ rr ] 1 2 E or E 1 [ + rr ] = [ rr ] = E 1 2

(4.3.3)

( rr + ), zr = z = 0 and zz = 0 . E

Figure 4.3.5: plane stress axisymmetric problem

In the plane strain case, the strains zz , z and zr are zero. This will occur, for example, in a hollow cylinder under internal pressure, with the ends fixed between immovable platens, Fig. 4.3.6.

Figure 4.3.6: plane strain axisymmetric problem

Solid Mechanics Part II

68

Kelly

Section 4.3

Hookes law 4.2.5b reads

rr

E 1 + [ + (1 ) rr ] = rr [(1 ) rr ] = ( )( ) 1 1 2 + E (4.3.4) or 1 + E [(1 ) rr ] = [ + (1 ) ] = E (1 + )(1 2 ) rr

with zz = ( rr + ) . Shown in Fig. 4.3.7 are the stresses acting in the axisymmetric plane body (with zz zero in the plane stress case).

zz rr

rr zz

Figure 4.3.7: stress components in plane axisymmetric problems

4.3.4

Solution of Plane Axisymmetric Problems

The equations governing the plane axisymmetric problem are the equations of equilibrium 4.2.3 which reduce to the single equation
rr 1 + ( rr ) = 0 , r r

(4.3.5)

the strain-displacement relations 4.3.2 and the stress-strain law 4.3.3-4. Taking the plane stress case, substituting 4.3.2 into the second of 4.3.3 and then substituting the result into 4.3.5 leads to (with a similar result for plane strain) d 2 u 1 du 1 + u=0 dr 2 r dr r 2 (4.3.6)

This is Naviers equation for plane axisymmetry. It is an Euler-type ordinary differential equation which can be solved exactly to get (see Appendix to this section, 4.3.8) u = C1 r + C 2 1 r (4.3.7)

Solid Mechanics Part II

69

Kelly

Section 4.3

With the displacement known, the stresses and strains can be evaluated, and the full solution is
u = C1 r + C 2 1 r

rr

1 1 , = C1 + C 2 2 2 r r 1 1 E E E E = C1 C 2 2 , = C1 + C2 2 1 1 + 1 1 + r r

rr = C1 C 2

(4.3.8)

For problems involving stress boundary conditions, it is best to have simpler expressions for the stress so, introducing new constants A = EC 2 / (1 + ) and C = EC1 / 2(1 ) , the solution can be re-written as 1 1 + 2C , = A 2 + 2C 2 r r (1 + )A 1 + 2(1 )C , = (1 + )A 1 + 2(1 )C , = 4C =+ (4.3.9) zz E E E E E r2 r2 (1 + )A 1 + 2(1 )C r u= E r E Plane stress axisymmetric solution

rr = + A

rr

Similarly, the plane strain solution turns out to be again 4.3.8a-b only the stresses are now {Problem 1}

rr =

E 1 E 1 + (1 2 )C 2 2 + C1 (1 2 )C 2 2 + C1 , = (1 + )(1 2 ) (1 + )(1 2 ) r r (4.3.10)

Then, with A = EC 2 / (1 + ) and C = EC1 / 2(1 + )(1 2 ) , the solution can be written as

rr

1 1 + 2C , = A 2 + 2C , zz = 4C 2 r r 1 + 1 1 + 1 = + A 2 + 2(1 2 )C , = A 2 + 2(1 2 )C (4.3.11) E E r r 1 + 1 u= A + 2(1 2 )Cr E r Plane strain axisymmetric solution

rr = + A

The solutions 4.3.9, 4.3.11 involve two constants. When there is a solid body with one boundary, A must be zero in order to ensure finite-valued stresses and strains; C can be determined from the boundary condition. When there are two boundaries, both A and C are determined from the boundary conditions.

Solid Mechanics Part II

70

Kelly

Section 4.3

4.3.5

Example: Expansion of a thick circular cylinder under internal pressure

Consider the problem of Fig. 4.3.8. The two unknown constants A and C are obtained from the boundary conditions

rr (a) = p rr (b) = 0
which lead to

(4.3.12)

rr (a) =
so that

A A + 2C = p, rr (b) = 2 + 2C = 0 2 a b

(4.3.13)

rr

b2 / r 2 1 b2 / r 2 +1 = p 2 2 , = + p 2 2 , zz = ( rr + ) (4.3.14) b / a 1 b / a 1 Cylinder under Internal Pressure

a r
b

Figure 4.3.8: an internally pressurised cylinder

The stresses through the thickness of the cylinder walls are shown in Fig. 4.3.9a. The maximum principal stress is the stress and this attains a maximum at the inner face. For this reason, internally pressurized vessels often fail there first, with microcracks perpendicular to the inner edge been driven by the tangential stress, as illustrated in Fig. 4.3.9b. Note that by setting b = a + t and taking the wall thickness to be very small, t , t 2 << a , and letting a = r , the solution 4.3.14 reduces to:

rr = p, = + p , zz = p

r t

r t

(4.3.15)

which is equivalent to the thin-walled pressure-vessel solution, Part I, 4.5.2 (if = 1 / 2 , i.e. incompressible).

Solid Mechanics Part II

71

Kelly

Section 4.3

p p

b2 / a 2 + 1 b2 / a 2 1
2

zz 2 2 b / a 1
p

rr
r =b (a )

r=a

( b)

Figure 4.3.9: (a) stresses in the thick-walled cylinder, (b) microcracks driven by tangential stress

Generalised Plane Strain Solution

Another useful solution is that for a cylinder which is free to expand in the axial direction. In this case, zz is not forced to zero as in plane strain, but allowed to be a constant along the length of the cylinder. The zz stress is zero, as in plane stress. This situation is called generalized plane strain. Returning to the full three-dimensional stress-strain equations (Part I, Eqns. 4.2.9), set zz = zz , a constant, and xz = yz = 0 . Re-labelling x, y, z with r , , z , and again with

r = 0 , one has rr = zz
E [(1 ) rr + ( + zz )] (1 + )(1 2 ) E [(1 ) + ( rr + zz )] = (1 + )(1 2 ) E = [(1 ) zz + ( rr + )] = 0 (1 + )(1 2 )

(4.3.16)

Substituting the strain-displacement relations 4.3.2 into 4.3.16a-b leads to

rr =
where

u r u + r + ( + ) zz r r u u = r + r + ( + ) zz r r

(4.3.17)

Solid Mechanics Part II

72

Kelly

Section 4.3

E (1 ) , (1 + )(1 2 )

E (1 + )(1 2 )

(4.3.18)

Substituting 4.3.16 into the axisymmetric equilibrium equation 4.3.5 again leads to the differential equation 4.3.6. The solution for displacement and strain is thus again 4.3.8ab. The constant axial strain is obtained from 4.3.16c and is zz = 2C1 / (1 ) . The axial displacement is then u z = z zz (to within a constant). The stresses are again given by the plane strain relations 4.3.10 only with the additional term 2 E 2 C1 / (1 2 )(1 2 ) . As for plane strain, let A = EC 2 / (1 + ) and C = EC1 / 2(1 + )(1 2 ) , and the full solution is
1 4 2 1 4 2 rr = + A 2 + 2C C , = A 2 + 2C C 1 1 r r 1 + 1 1 + 1 = + A 2 + 2(1 2 )C , = A 2 + 2(1 2 )C (4.3.19) E E r r 1 + 1 A + 2(1 2 )Cr u= E r Generalised plane strain axisymmetric solution

rr

A Transversely isotropic Cylinder

Consider now a transversely isotropic cylinder. The strain-displacement relations 4.3.2 and the equilibrium equation 4.3.5 are applicable to any type of material. The stressstrain law can be expressed as (see Part I, Eqn. 6.2.14)

rr = C11 rr + C12 + C13 zz = C12 rr + C11 + C13 zz zz = C13 rr + C13 + C 33 zz

(4.3.20)

Here, take zz = zz , a constant. Then, using the strain-displacement relations and the equilibrium equation, one again arrives at the differential equation 4.3.6 so the solution for displacement and strain is again 4.3.8a-b. With A = C 2 / (C12 C11 ) and C = C1 / 2(C11 + C12 ) , the stresses can be expressed as

rr = + A zz

1 + 2C + C13 zz r2 1 = A 2 + 2C + C13 zz r C13 = 4C + C 33 zz C11 + C12

(4.3.21)

The plane strain solution then follows from zz = 0 and the generalized plane strain solution from zz = 0 . These solutions reduce to 4.3.11, 4.3.19 in the isotropic case.
Solid Mechanics Part II 73 Kelly

Section 4.3

4.3.6

Stress Function Solution

An alternative solution procedure for axisymmetric problems is the stress function approach. To this end, first specialise equations 4.2.6 to the axisymmetric case:

rr =

1 2 , = 2 , r = 0 r r r

(4.3.22)

One can check that these equations satisfy the axisymmetric equilibrium equation 4.3.4. The biharmonic equation in polar coordinates is given by Eqn. 4.2.7. Specialising this to the axisymmetric case, that is, setting / = 0 , leads to

2 1 2 1 2 1 + = r 2 r r r 2 + r r 2 + r r =0 r or d 4 2 d 3 1 d 2 1 d + + =0 dr 4 r dr 3 r 2 dr 2 r 3 dr

(4.3.23)

(4.3.24)

Alternatively, one could have started with the compatibility relation 4.2.8, specialised that to the axisymmetric case: 2 1 rr 2 + =0 r r r r r 2 and then combine with Hookes law 4.3.3 or 4.3.4, and 4.3.22, to again get 4.3.24. Eqn. 4.3.24 is an Euler-type ODE and has solution (see Appendix to this section, 4.3.8)

(4.3.25)

= A ln r + Br 2 ln r + Cr 2 + D
The stresses then follow from 4.3.22:

(4.3.26)

rr = +

A + B(1 + 2 ln r ) + 2C r2 A = 2 + B(3 + 2 ln r ) + 2C r

(4.3.27)

The strains are obtained from the stress-strain relations. For plane strain, one has, from 4.3.4,

Solid Mechanics Part II

74

Kelly

Section 4.3

rr =

1 + A + 2 + B[1 4 + 2(1 2 ) ln r ] + 2C (1 2 ) E r 1 + A = 2 + B[3 4 + 2(1 2 ) ln r ] + 2C (1 2 ) E r

(4.3.28)

Comparing these with the strain-displacement relations 4.3.2, and integrating rr , one has u r = rr dr = u r = r A + Br [ 1 + 2(1 2 ) ln r ] + 2C (1 2 )r + F r 1 + A = + Br [+ 1 + 2(1 2 ) ln r ] + 2C (1 2 )r E r 1 + E

(4.3.27)

To ensure that one has a unique displacement u r , one must have B = 0 and the constant of integration F = 0 , and so one again has the solution 4.3.113.

4.3.7

Problems

1. Derive the solution equations 4.3.11 for axisymmetric plane strain. 2. A cylindrical rock specimen is subjected to a pressure p over its cylindrical face and is constrained in the axial direction. What are the stresses, including the axial stress, in the specimen? What are the displacements? 3. A long hollow tube is subjected to internal pressure pi and external pressures p o and constrained in the axial direction. What is the stress state in the walls of the tube? What if pi = p o = p ? 4. A long mine tunnel of radius a is cut in deep rock. Before the mine is constructed the rock is under a uniform pressure p. Considering the rock to be an infinite, homogeneous elastic medium with elastic constants E and , determine the radial displacement at the surface of the tunnel due to the excavation. What radial stress rr (a) = P should be applied to the wall of the tunnel to prevent any such displacement? 5. A long hollow elastic tube is fitted to an inner rigid (immovable) shaft. The tube is perfectly bonded to the shaft. An external pressure p is applied to the tube. What are the stresses and strains in the tube?

the biharmonic equation was derived using the expression for compatibility of strains (4.3.23 being the axisymmetric version). In simply connected domains, i.e. bodies without holes, compatibility is assured (and indeed A and B must be zero in 4.3.26 to ensure finite strains). In multiply connected domains, however, for example the hollow cylinder, the compatibility condition is necessary but not sufficient to ensure compatible strains (see, for example, Shames and Cozzarelli (1997)), and this is why compatibility of strains must be explicitly enforced as in 4.3.25

Solid Mechanics Part II

75

Kelly

Section 4.3

6. Repeat Problem 3 for the case when the tube is free to expand in the axial direction. How much does the tube expand in the axial direction (take u z = 0 at z = 0 )?

4.3.8

Appendix

Solution to Eqn. 4.3.6

The differential equation 4.3.6 can be solved by a change of variable r = e t , so that

r = e t , log r = t , and, using the chain rule,

1 dt = r dr

(4.3.28)

du du dt 1 du = = dr dt dr r dt 1 d 2 u 1 du d 2 u d du dt d 2 u dt dt du d 2 t = = + = dr 2 dr dt dr dt 2 dr dr dt dr 2 r 2 dt 2 r 2 dt

(4.3.29)

The differential equation becomes d 2u u = 0 dt 2 (4.3.30)

which is an ordinary differential equation with constant coefficients. With u = e t , one has the characteristic equation 2 1 = 0 and hence the solution
u = C1e + t + C 2 e t = C1 r + C 2 1 r

(4.3.31)

Solution to Eqn. 4.3.24

The solution procedure for 4.3.24 is similar to that given above for 4.3.6. Using the substitution r = e t leads to the differential equation with constant coefficients d 4 d 3 d 2 4 + 4 =0 dt 4 dt 3 dt 2
2

(4.3.32)

which, with = e t , has the characteristic equation 2 ( 2) = 0 . This gives the repeated roots solution

= At + Bte 2t + Ce 2t + D
and hence 4.3.24.
Solid Mechanics Part II 76

(4.3.33)

Kelly